Decart raises 300 million dollars in Nvidia and Amazon backed round

Yesterday 09:38

By: Dakir Madiha

Decart raises 300 million dollars in Nvidia and Amazon backed round

Decart, a San Francisco based artificial intelligence research lab focused on real time world models, has raised 300 million dollars in a new funding round led by Radical Ventures. The deal brings the company’s total funding to more than 450 million dollars and values it at around 4 billion dollars, up from 3.1 billion dollars in August 2025.

The financing round includes major technology and investment players. Nvidia joins as both investor and strategic partner, alongside Amazon, Atreides Management, Valor Equity Partners, Adobe Ventures, Toyota Ventures and eBay Ventures. Existing backers such as Sequoia Capital, Benchmark and Zeev Ventures also participated, reinforcing strong institutional support around the company’s long term ambitions in artificial intelligence infrastructure and world modeling systems.

Decart has also attracted a high profile group of angel investors from the technology and entertainment sectors. Among them are OpenAI co founder Andrej Karpathy, former Disney chief executive Michael Eisner, members of the Nintendo family and gaming investor Moritz Baier Lentz, reflecting growing cross industry interest in real time AI systems that can simulate physical and digital environments.

The company develops three core products. DOS is an optimization stack designed for AI training and inference workloads, with reported performance of up to 1,600 tokens per second for agent based systems. Lucy is a real time world model designed for interactive environments, capable of responding to user input in under 30 milliseconds. Oasis focuses on physical world simulation for robotics and autonomous systems, positioning Decart in the emerging field of embodied artificial intelligence.

A key commercial element of the deal is Decart’s integration with Amazon Web Services infrastructure. The company is among the first to deploy large scale real time AI models on AWS Trainium hardware, with its Lucy2 model running on Trainium3 chips. Amazon’s Annapurna Labs leadership said the system achieves more than 80 percent hardware utilization, highlighting efficiency gains in specialized AI workloads.

Nvidia’s participation reflects a broader investment strategy that has seen the company deploy tens of billions of dollars into AI related startups. Decart’s software stack is notable for operating across Nvidia GPUs, Google TPUs and Amazon Trainium chips, positioning the company as a rare cross platform benchmark in the competitive AI hardware ecosystem.

Decart chief executive Dean Leitersdorf described world models as essential for bridging the gap between virtual intelligence systems and physical reality, arguing that current language models lack a true understanding of how the physical world operates. The company announced DOS 2.0 alongside the funding round, with upcoming releases of Lucy 2.5 and Oasis 3 expected in the coming weeks.